Performance of novices in Endoscopic Submucosal Dissection starting directly in human patients under direct supervision of an expert endoscopist.

Background and study aim: The ESGE curriculum guideline on training in ESD advises an extensive experience in animal models before commencing ESD training on humans. However, the evidence on which this recommendation is based, is rather limited. In this study we evaluate the performance of novices in ESD during their one-year training period; with continuous supervision of an endoscopist with extensive experience in ESD.

Patients and methods: During a 6.5 years period, all ESD procedures performed by an expert endoscopist (eESD) or by one of the 3 novices (sESD), in a single non-academic center, were retrospectively collected. Primary outcome parameters were rate of en-bloc resection and clinically relevant adverse events. The secondary outcome was R0 resection, recurrence rate and delayed adverse events.

Results: In total 210 ESD procedures were included. Of these 97 sESD (46.2%), were performed by a novice endoscopist under supervision and 113 eESD (53,8%) by an expert. En-bloc resection rate was 99.1% vs 99% (p = ns), R0 resection rate was 80.5 % vs 79.4 % (p=ns). Clinically relevant complications 14.2% vs 6.2 % (p = 0.04) and severe complications 4.4% vs 2.1% (p=ns). Perforation was the most frequent reported serious complication. Three patients in the eESD group versus one in the other group needed additionally surgery due to a complication (p=ns). Recurrence rate was 2.2% for eESD and 1.3 % for sESD (p=ns).

Conclusion: This retrospective analysis demonstrates that learning ESD in humans without extensive prior animal model training is safe and feasible, on the condition that continuous supervision by an expert is provided.